27-09-2019

Here’s how IFFCO Tokio General Insurance has automated claim settlement with AI

Insurance Alertss
|
27-09-2019
|

Here’s how IFFCO Tokio General Insurance has automated claim settlement with AI

Technological advancements and changing customer preferences have compelled insurers to adopt various models to reach out to all industry segments. Being in the business for the past 19 years, IFFCO Tokio General Insurance has adopted an omnichannel distribution strategy, propelled by a strong IT backbone, to reach different customer segments as per their convenience. As a result, the company has an omnipresence in all digital platforms, including portal and mobile device which are assisted with customer support through chat and IVR.
Seema Gaur, ED & Head- IT, IFFCO Tokio General Insurance, has recently been recognised by IBM in 40 Women Business Pioneers list across the globe in AI innovations. She has been successfully leveraging AI in helping customers with quicker claim process in vehicle damage.
ETCIO caught-up with her to get insights into IFFCO Tokio’s automation journey and how she settled 33% of personal car claims through AI.
How are you leveraging AI in motor insurance claims?
We leverage AI for image processing to analyse the extent of the damaged vehicle (personal cars), which helps us to generate a list of repairable and replaceable parts that were damaged in the accident. Within a few minutes, the assessment and the cost are given to the customer, which they can either accept or reject. If the assessment gets accepted, the payment is made within fifteen minutes directly to the customer’s bank account.
In the absence of AI, these photographs were assessed by claim officers, and it was a tedious and manual jobs, as they took a few hours to assess the damage. With the AI-based app, the list of damaged parts is generated within minutes along with a cost. It is now much more convenient for claim officers, as they are now in a better position to justify the assessment to the claimant.
To ensure that there is no error, the assessment data is also analysed by human intervention in the back office. If any error is found, the AI model is trained through feedback mechanism, as proper data input is one of the essential requirements of machine learning. Considering, it is a logical way of assessment, our acceptability of claim estimate has risen to over 75%. Moreover, the accuracy of the model has gone up beyond 80-85%. Currently, out of all the QCS (Quick Claim Settlement) that we settle for personal cars 33% are through AI.
What were the challenges you faced while working on this project?
There are a few challenges we faced. For example, we have started with 30 top and popular models of cars. We will be adding another 20 models soon. The other challenge is that it is limited to exterior parts of the vehicle. There are certain parts which cannot be apparently seen as damaged, as they may be parent and child parts which need to be repaired together.
If the child part gets damaged the parent part also has to be replaced or vice versa. For better assessment, we took the services of automobile experts, from within the company, to arrive at correct estimates. There are also parts for which historical data is not available, as various manufacturers keep on upgrading their models frequently. This sometimes might result in delay in assessment.
Going forward, we plan to use AI for pre-inspection where break-ins occur. Moreover, we plan to bring more vehicle types, as currently AI is used only for personal cars and not for commercial vehicles and two-wheelers.
What is the use case for RPA in IFFCO Tokio?
We have been investing in RPA (Robotic Process Automation) for our in-house surveyors. RPA is deployed for automated entry of motor claim invoices into the system. We rolled out the facility last month after we ran a pilot for two months. We have done 7500 entries into the system so far at an average of 250 invoices per day. Our target is to increase it by 3X to reap the benefits. The initiative has helped us in bringing the cost down by saving on significant number of man-hours without compromising on our compliance standards.
We are also trying to deploy RPA in other repetitive or routine activities which are used by our supporting staff. The automation has helped us in bringing process optimisation and enhanced customer service. We are also working on AI/ML based fraud prevention and detection tool for our motor and health claims.
How are you managing IT security?
We have an IT infrastructure that supports continued growth. We are accredited by ISO 27000: 2013, which is an industry standard for Information Security Management System (ISMS). We have taken strict measures for the security of critical data stored in our systems. For example, we ensure data transmission flows in a secure encrypted manner through SSL (Secure Socket Layer). Similarly, all our portals meant for intermediaries and customers go through a stringent VAPT (Vulnerability Assessment and Penetration Test) twice a year, to plug any security loopholes in the system.
For the security of our critical data, we have got firewalls, for data centre security in the inner and outer periphery. We also use different technologies, such as DLP (Data Leak Prevention) and endpoint security, for protecting and controlling confidential data flow outside the system. We also have devices to protect our network from brute force attack like DDOS (Distributed Denial of Service).
How are you leveraging cloud?
Cloud computing has changed the landscape and insurers also embraced it to save cost and enhance efficiency. All our mobile applications meant for intermediaries and customers are now deployed through the cloud. Our website is also hosted on the cloud and for that, we have engaged more than one cloud service provider. Moreover, we ensure that data flow happens through an express route between the cloud hosting site and our Data Center. Moreover, to ensure business continuity we have the near line as well as far site disaster recovery centres. All our transactions are replicated in real-time at these sites. We also conduct DR-Mock drills annually.
Talking about the Blockchain consortium, what are the benefits of this open-model and key initiatives of the group?
It is a new venture started by life and non-life companies. The non-life consortium was started one and a half years back. We have been using this for NCB verification. The purpose is to verify that a particular vehicle registration number has got a claim in the previous year, as it can have an impact on NCB. As of now, only a few companies have participated in the consortium. We are trying to rope in more Insurers in it.
Source: The Economic Times